Writing Around the Ancient Mediterranean: Practices and Adaptations

By Oxbow Books

Writing in the ancient Mediterranean existed against a backdrop of very high levels of interaction and contact. In the societies around its shores, writing was a dynamic practice that could serve many purposes – from a tool used by elites to control resources and establish their power bases to a symbol of local identity and a means of conveying complex information and ideas. This volume presents a group of papers by members of the Contexts of and Relations between Early Writing Systems (CREWS) research team and visiting fellows, offering a range of different perspectives and approaches to problems of writing in the ancient Mediterranean. They focus on practices, viewing writing as something that people do within a wider social and cultural context, and on adaptations, considering the ways in which writing changed and was changed by the people using it.

• Language: English
• Publisher: Oxbow Books
• Release date: Oct 6, 2022
• ISBN: 9781789258516

Book preview

Chapter 1

Introduction: approaches to the study of writing, and the development of the CREWS project

Philippa M. Steele

I hope the reader will indulge me in writing this perhaps unusual introductory chapter, which will focus on contextualising the present conference volume as a contribution to the research output of the CREWS project (Contexts of and Relations between Early Writing Systems). I suppose most readers will pass it by, especially those who have come to the volume following a trail of references to track down just one of the papers – that is the fate of most conference volumes eventually, after all. However, as the CREWS project draws to a close, with this as its penultimate published volume, I hope it will be of interest to some to offer a few reflections on what CREWS has been about and what we have tried to contribute to the study of writing. Since this conference was always intended as a culmination of the CREWS project, at which we would present the research of the core CREWS research team and wider CREWS family, this is also a useful opportunity to explain how the papers in this volume and their authors have fitted into (and in other ways expanded) the project’s remit.

Back in 2015 I submitted an application for funding to bring a quite abstract idea in my head into reality, an idea revolving around the ways I had originally learnt to talk about writing systems and the quite different ways I had started to think about writing systems as I began to conduct my own research. I was used to the old narrative that grouped writing systems in terms of their linguistic properties, labelling them as alphabets if they encoded each sound with a separate sign, syllabaries if they encoded whole syllables, abugidas or semi-syllabaries if they did something between the two, and so on. These are helpful designations to be sure, but it often felt that the people doing the writing and the place of writing activities in society had gone missing somewhere. Can you really understand the way a writing system represents language completely in isolation from questions surrounding its users and usage? A second problem also preyed on my mind, after spending many years working on the syllabic systems of the Bronze Age Aegean and Cyprus: what exactly do we mean when we say that two writing systems are related to each other? We might sometimes talk about ‘families’ of writing systems, but is the quasi-Darwinian terminology we apply in historical linguistics (language families, ancestors and descendants, etc.) really appropriate here? Writing system developments are the product of complex and socially contextualised motivations and changes, and there are no ‘sign laws’ or ‘letter laws’ in the same way that we assume there are sound laws, no regularity in writing system changes such as that hypothesised for sound changes. But if those traditional approaches originating from linguistic research are not the best ways of approaching such questions, then what methods should we use? And what about other kinds of interactions between writing systems and the people and societies using them? And on the back of these sorts of questions, the idea for the CREWS project came about. The project was funded in 2016, and over the last six years has developed in some ways that I had hoped for, and a whole range of other ways I had not expected.

Contexts and relations, practices and adaptations

I have to admit that the choice of the words ‘practices’ and ‘adaptations’ in the title of this book and the conference preceding it may not have been entirely unrelated to the desire to find a nice acronym. But more seriously, I was thinking of ‘practices and adaptations’ as an interesting proxy for the words ‘contexts and relations’ that have been enshrined in the title of the CREWS project from the beginning. For example, studying context very often involves thinking about the ways and places in which something was done, i.e. its practice, and writing systems and practices tend to undergo adaptation both at the point of their borrowing from one group by another and throughout their lifetimes, creating relationships between different traditions in different situations.

We have talked in more detail in the previous CREWS conference volume about some of the different ways in which writing can be approached, whether the perspectives are structuralist or culturally driven, or whether they stem from linguistic, (social-)archaeological, anthropological or other disciplinary backgrounds (Boyes et al. 2021). Those reflections stemmed from our efforts both to try to approach writing from new and interesting angles, and to bring some of those different angles together to gain deeper insights into the way writing works. Some of those efforts originate from the direct research of the core CREWS research team (on which more below), but they were also bolstered and inspired by interactions with other scholars working on a range of ancient writing systems, including CREWS Visiting Fellows (the ‘CREWS family’, many of whom contributed to this volume) and other friends and colleagues who participated in our conferences and other events.

The core CREWS research team consists of four researchers, each pursuing a different case study situated somewhere around the eastern half of the Mediterranean: Natalia Elvira Astoreca, Philip Boyes, Robert Crellin and me as PI. Natalia took a novel, graphematic approach to the diversity of the epichoric Greek alphabets of the Archaic period, an archetype example of a set of systems whose properties have usually been discussed as if the individual systems had diverged from an unattested common ancestor; she viewed their differences not in terms of their palaeography (which has been the focus of most previous studies) but rather as a series of potential different solutions to linguistic problems, teasing apart different motivations for different types of variation and presenting what have traditionally been referred to as ‘local scripts’ as fully independent writing systems (Elvira Astoreca 2021). Philip took a social-archaeological approach to the practice of writing at Bronze Age Ugarit, seeing writing as intrinsically and dynamically bound up with a whole range of other ideas and practices, and establishing how we might pursue more holistically an ‘archaeology of writing’; by studying the wider social and political context of writing, this made it possible to explore the people and places involved in its practice (Boyes 2021a). Robert, following a first stage of research on the phenomenon of vowel writing in different systems (e.g. Crellin 2020 focused on Neo-Punic), contributed a monograph comparing the practice of word division in Greek and Northwest Semitic writing and its relationship with prosody and orality; although ostensibly dealing with linguistic analysis of orthography, the results help us to understand the design and context of these writing traditions as they point strongly towards features being motivated by a desire to make what is written easier to read aloud (Crellin 2022). My own research as PI has drawn together some of these strands, among others, to attempt to establish an integrated approach to writing, using the syllabic systems of the Bronze Age Aegean as a test case (Steele forthcoming, presaged in Steele 2020) – since this research is not yet published, I won’t venture to comment on what it may contribute to the field of writing studies just yet.

The combination of the specificity of the case studies, alongside the much wider applicability of the methodological approaches pursued, matches the original aims of the CREWS project well: to contribute to the fields of research on particular ancient writing systems while pursuing methodological innovation that helps us to rethink the ways writing works and the ways we talk about it. Along the way this involved a lot of group discussion, bringing together different material and different approaches and trying to identify overlapping areas of interest and common ground, as well as having fun with some very revealing practical experiments that helped us to understand the experiences of ancient writers. Rather than pointing towards commonalities between different writing traditions, this helped us to appreciate the extent to which writing in any one society can only be understood fully on its own terms and within its own unique context.

The CREWS Visiting Fellowship scheme gave an opportunity to expand the CREWS remit both in terms of methodology and especially in terms of the range of writing systems and societies in focus. Most Visiting Fellows have contributed a major piece of their CREWS-related research to this volume, on which see further below. However, there were some colleagues from the CREWS family who were unable to contribute to the conference and/or this volume: Brent Davis (who had visited in 2020 while working on syllabotactic analysis of the Aegean linear scripts), Giorgos Bourogiannis (whose own project, Cypriot Connectivity in the Mediterranean from the Late Bronze Age to the end of the Classical Period [CyCoMed], has been closely affiliated with CREWS) and Kathryn Piquette (who gave us a conference paper on her research on embodied practice and early Egyptian writing).¹ Our previous two conferences also expanded our outlook considerably and were an only partially expected source of inspiration for our core research: Understanding Relations Between Scripts II: Early Alphabets² (Boyes and Steele 2019) and The Social and Cultural Contexts of Historic Writing Practices (Boyes et al. 2021).

This volume and the CREWS family

The conference title, Writing around the Ancient Mediterranean: Practices and Adaptations, was not intended to be prescriptive, but it does encapsulate quite nicely how the research of the CREWS core team and wider family has revolved around questions of the contexts and relatedness of writing systems and traditions.

We begin with Csaba La’da’s paper, demonstrating that a writing system can also act as a sort of cultural artefact, taking on uses beyond the notation of language – most obvious in the various alphabetic systems that have a fixed order of signs, and so have structural properties that can be lent to other areas of society. Such ‘secondary uses’ of writing include numerical notation, other ways of encoding and decoding information, musical notation and, perhaps the most prevalent and widely successful to this day, the ordering of information by alphabetical order (which was indeed the prompt for this paper in looking at some of the earliest examples of the phenomenon). Michel de Vreeze then looks at some of the more short-lived alphabetic traditions of the Levant, focusing on what we can reconstruct of their usage to try to understand their initial vitality and then their ensuing loss. He argues for relatively restricted contexts of use, particularly ritual ones, and considers how they were negotiated against a background of other social phenomena and wider developments across the region.

Although Cyprus was never specifically intended to provide case studies for the CREWS project (particularly given that I have already written rather a lot about it!), this materially rich island at the heart of the eastern Mediterranean has come up again and again, and with good reason. The next chapter, by Cassandra Donnelly,³ focuses on the elusive single-sign inscriptions of Bronze Age Cyprus, usually dismissed as ‘potmarks’ or similar that do not count as proper inscriptions, and their relationship with two-sign inscriptions more usually included in the corpora. While the single sign inscriptions may present numerous problems in any attempt to incorporate them into our understanding of the Cypro-Minoan writing system and its repertoire, they do however help us to understand writing practices and especially the use of acrophonic abbreviation more widely. Martina Polig’s chapter also looks at the practice of Bronze Age Cypriot writing, approached through a groundbreaking 3D documentation of sign shapes. By looking at the minute details of sign composition, she shows that sign shapes and ductus are closely linked with networks of writing practice that include elements of design and embodied tool use – allowing observations that are all the more important for a writing system whose exact repertoire (in terms of both individual signs and their variation in shape) remains only partially established.

The chapter by Philip Boyes takes a trans-regional approach, looking at the relationship between writing and the practice of magic across Egypt and the Near East, where we see some cultural influence and exchange in operation, as well as similar phenomena that may have arisen independently. Rather than seeing writing as having an inherent mysticality, especially to the illiterate, this considers writing as an important component of magical practice in different societies. Willemijn Waal then offers a chapter on writing in Bronze Age Anatolia, where a localised ‘hieroglyphic’ writing system co-existed for some time with Hittite cuneiform but was used largely in different spheres. She uses aspects of the arrangement of Anatolian hieroglyphic text, alongside other evidence, to reconstruct a writing tradition on wood for which direct attestations have inevitably been lost, positing reflections of this tradition in writing on other media such as stone.

Moving to the first millennium BC, Rostislav Oreshko’s chapter looks at the rarer letters of the Phrygian alphabet and the puzzle of their phonetic values and their origins. This involves consideration of not only the other Anatolian alphabets but also the closely related Greek alphabet, shedding light on their different paths of development. The Archaic regional Greek alphabets are also the subject of Natalia Elvira Astoreca’s paper, which uses the computational methodology of Natural Language Processing (NLP) to measure the similarities and differences between them. Despite the challenges of a relatively small and imbalanced dataset, these methods offer the advantage of processing data in such a way as to point out patterns that have not previously been identified.

Beatrice Pestarino’s paper again focuses on ancient Cyprus, this time looking at the first millennium BC and the late arrival of the Greek alphabet on an island whose writing traditions were dominated by the local syllabic system. She sees the earliest attestations of alphabetic writing in the sixth century BC as being bound up with the attempts of Cypriot elites to harness aspects of visual culture, aiming to establish their authority through connections with the wider Greek-speaking world. Robert Crellin then uses another cosmopolitan island, Sicily, to investigate the development and types of word division in Latin and Greek. The evidence from the Imperial era shows prosodic word division in Greek continuing later than expected, with signs of influence and exchange with word division in Latin.

Perhaps the most striking aspect of all these papers is that the concepts of adaptation and practice are never encountered entirely separately from each other; they are very much bound up in the same activities and developments as writing traditions in and across different societies grow. Much the same could be said of the concepts of the ‘contexts’ and ‘relations’ embedded in the CREWS project remit, which may have begun as separate sets of research questions but soon turned out to be contributing to a much more integrated outlook on how writing works and how it developed in ancient societies.

The final paper in the volume is something of a departure from the others. Charles ‘Pico’ Rickleton is an art director and designer, who contacted me some time ago with an idea that was to have a considerable influence both on the CREWS project and on my own research. His starting question was simple: what if the Cypriot syllabic writing system, so characteristic of Cypriot writing during most of the first millennium BC, had survived to the present day? After spending a great deal of my research career thinking about this writing system and the way it dwindled and disappeared from the archaeological record during the Hellenistic period, the question rather blew my mind. And so began an exciting collaboration, which invited reflection on why and how writing systems maintain vitality, and the possible alternative universe conditions in which Cypriot syllabic writing might have survived and subsequently changed over the last 2,000 years. I hope that Pico’s paper, and more generally his ‘Speculative Syllabic’ project, will be as stimulating and exciting to others as it has been to me.

The CREWS legacy

As the CREWS project draws to a close, I would like to think that it has done something useful in the world of writing systems research, and perhaps more generally. It will be for others to judge the project’s success and the effectiveness of our research will only unfold (I hope!) as time goes on and the wider field engages with it.⁴ But I will finish this (already slightly self-indulgent) chapter by mentioning some of my own favourite legacies of the project.

The CREWS family of Visiting Fellows, has itself become symbolic of the spirit of collaboration that I hope has marked the operation of the project. It has been a huge pleasure to welcome this group of individuals to Cambridge, with their different research areas and yet often similar or overlapping interests, and it has been highly stimulating to engage in discussion with them. I have learnt so much from our conversations.

Fig. 1.1. Clockwise from top left: Illustration from Writing in the Ancient World, created for us by Katie Idle; one of Pippa’s experiments with writing Cypro-Minoan on modelling clay; the CREWS display case at the Fitzwilliam Museum in 2018; Philip’s (edible and delicious) Phaistos Discuits.

The CREWS blog website (along with our Twitter feed) has proved to be very enjoyable to run, especially in the way it has engendered interaction with a much wider group of interested individuals, some in academia and many outside of it.⁵ I am particularly pleased with the interest in following some of our practical experiments, which have been an important feature of CREWS research from the beginning, and which have occasionally produced edible results.

Outreach has always been an important aspect of CREWS, and over the years we have been involved in many different events both in person and online, producing materials and blog content that we hope will help the wider public to engage with the ancient societies and writing systems we work on. This has included writing a lot of blog posts, playing with quite a bit of Lego (in the name of research!), producing worksheets to show people how to write their name in an ancient writing system and exploring links between what we do and other areas of popular culture. In 2018 we also collaborated with the Fitzwilliam Museum and British Museum (with thanks to Anastasia Christophilopoulou and Thomas Kiely) to put on a display of ancient writing systems in the Fitzwilliam’s Leventis Gallery.

Stemming from all our outreach work, in 2020 we worked on a pedagogical initiative, Writing in the Ancient World, which involved the creation of free videos and teaching materials aimed primarily at children aged 7–11 (for which most of the hard work was done by Philip).⁶ The feedback we have had from this initiative has been a joy to receive, showing the CREWS project making its mark in all sorts of unexpected ways on people across the world.

It only remains for me to thank everyone who has been involved in the CREWS project at all levels for making the last six years so fruitful, stimulating and, above all, so much fun.

¹  The video of Kathryn’s conference paper can be viewed on our YouTube channel here: https://youtu.be/2_Y2xiLSszQ.

²  You may be wondering why this is a volume II. It was a thematic sequel to an earlier volume, Understanding Relations Between Scripts: The Aegean Writing Systems (published with Oxbow in 2017), which arose from a 2015 conference before the CREWS project started up.

³  Cassie also worked on another topic during her visit in 2019, namely the inscriptions found on bronze bowls of the Late Bronze to Early Iron Age around the eastern Mediterranean, again a subject with Cyprus at its heart.

⁴  CREWS publications have all been released with open access. You can find them here: https://crewsproject.wordpress.com/crews-publications/

⁵  https://crewsproject.wordpress.com/, https://twitter.com/crewsproject

⁶  Writing in the Ancient World received some extra funding from the University of Cambridge’s Arts and Humanities Impact Fund. You can access the free materials here: https://crewsproject.wordpress.com/writing-in-the-ancient-world/

Chapter 2

What is an alphabet good for?

¹,²

Csaba A. La’da

When we talk about alphabetic writing, we tend to think immediately of its primary function: recording the spoken language phonetically, using the letters of the alphabet. This is of course understandable as it was the first and main aim of developing this type of writing. Recording the spoken language in a material form that can be transmitted both spatially and chronologically is the fundamental function of any writing system whatever its underlying system. However, we should not forget that writing could be used for other purposes as well, which over time assumed increasing importance even if not rivalling its first and foremost function. The alphabetic script was particularly suitable for being used for such other, I would call secondary, purposes and so in this regard too the idea of alphabetic writing had important consequences for intellectual developments in the ancient eastern Mediterranean.

What are these secondary functions of ancient writing, particularly of alphabetic scripts, that we see in the sources available to us currently? Without any intention of exhaustiveness,³ a quick count of the most important of these secondary functions of the scripts of the ancient eastern Mediterranean basin comes to five. However, it is important to bear in mind that not every script demonstrates all five (or more) secondary functions: some writing systems show some while other scripts demonstrate other such secondary functions. In this paper I intend to offer a brief survey, exempli gratia, of these various secondary functions and then to concentrate on one of them, the one that forms the subject of my own research.

Let us now look at these secondary functions briefly one by one. The order in which I shall discuss these is to some extent arbitrary and is in no way a reflection of the chronological order in which they arose or of any other, such as causal, relationship between them.

Numerals

The first such secondary function I would like to mention is that of using the elements of the alphabet as numerals. This idea may seem alien to us who have been accustomed to using separate characters for numbers for many centuries but this function of writing was widespread in the ancient Mediterranean world. Two main such systems developed across time and cultures: the first is generally called acrophonic and the second alphabetic.

The best-known example for such numeral systems is that of ancient Greek, in which both the acrophonic and the alphabetic systems were used. The acrophonic system in Greek, which is also called Herodianic after the second-century AD grammarian Herodianus who described it, is mainly attested in Attica, and was used in Greece from the seventh century BC until the beginning of the Roman Imperial period (in Attica from the mid-fifth century until about 95–90 BC). From the fifth century BC onwards it was gradually replaced by the alphabetic numeral system, which essentially supplanted it in the Greek world by the Roman Imperial period although there are sporadic examples of its survival from as late as the first and second centuries AD. The acrophonic system consisted of six simple characters for 1, 5, 10, 100, 1000 and 10000 and four compound signs. As the system’s name itself says, the numerals are actually the first letters of their Greek names, except for the unit, which was a straight vertical line. Thus, delta represented 10 from the word deka and chi 1,000 from chilioi.

The other main numeral system that enjoyed widespread popularity in the Greek world and that eventually became the most widely used is the alphabetic system. It essentially means that each letter of the alphabet is given a numerical value in alphabetic order. Thus, alpha equals 1, beta equals 2, gamma equals 3 and so on.⁴ The 24 letters of the classical Ionian alphabet were supplemented by three archaic letters, all ultimately derived from the Phoenician alphabet: stigma (or digamma) with the value of 6, qoppa with the value of 90 and sampi with that of 900. Thus, the full system consisted of 27 letters.

The alphabetic numeral system had the advantage of being economical, needing a maximum of three characters to represent any number under 1000,⁵ whereas the acrophonic system, albeit using fewer characters, needed several more. The alphabetic system became the dominant numeral system in Greek from the Hellenistic period onwards and was used all over the Greek-speaking world that, thanks to Alexander the Great’s conquest, came to include the whole of the eastern Mediterranean, Egypt and the Near and Middle East as far east as Afghanistan and northwest India. Its use was continued by the Greek-speaking administration of the eastern half of the Roman Empire that came to replace the Hellenistic states in many of these regions. Thus, the rich papyrological sources from Hellenistic, Roman and Byzantine Egypt supply plentiful evidence for the widespread use of the alphabetic numeral system for over a thousand years (Tod 1911–12; 1926–27; 1936–37; 1950; Dow 1952; Threatte 1980, 110–117; 1996, 278; Cook 1990, 267–268; McLean 2002, 58–64; Folkerts 2002, 670–676 [= 2006, 887–889], with the earlier works cited in these publications).

Where did the idea of using the letters of the alphabet for designating numerals come from? The alphabetic system of numerical notation clearly relies on a fixed order of the letters. This fixed order must be widely known to those able to read, write and count, otherwise the system simply cannot function. As is well known, the Greeks originally borrowed and later adapted the idea of alphabetic writing, as well as most letter forms, letter names and the order of the letters from the Phoenicians. Thus, it seems tempting to assume that the idea of using the letters of the alphabet for numerical notation in Greek also goes back originally to the same Northwest Semitic practice possibly via Miletus where the Greek system may have been developed (Swiggers 1996, 261; Folkerts 2002, 675 [= 2006, 888]; Paz and Weiss 2015, 54). It is theoretically possible that the Greeks adopted the idea of using letters of the alphabet for numerical notation at the same time they borrowed the idea of alphabetic writing as part of the same package, so to speak, including the basic idea of alphabetic writing, letter forms and names and their fixed order as well as the secondary function of the letters as numerals.

It also seems possible, however, that this secondary function is the result of a subsequent borrowing by the Greeks from the Phoenicians, which could be supported by the fact that the alphabetic system of numerical notation seems to appear and becomes dominant in Greek later than the acrophonic system. The long-standing contacts between the Greeks and Phoenicians in the eastern Mediterranean would easily allow such a secondary borrowing to take place in a later, but probably still pre-Classical, period (cf. Tod 1950, 137–138; Threatte 1980, 117 on a Locrian inscription of 460 BC). Whenever we hypothesise the transmission to have taken place, it is clearly a testimony to the closeness of cultural contacts between Greece and the Levant.

An alternative scholarly theory argues the exact opposite, namely that the alphabetic system of numerical notation was in fact invented in Greece, probably in Miletus before the end of the eighth century BC, and that this Greek invention then spread to the Levant and other parts of the Near East (Larfeld 1914, 293–297; Najock 1975; Healey 1990, 60; Pettersson 1996, 803).

Needless to say, the non-alphabetic writing systems of the eastern Mediterranean basin were less suitable for such a secondary use of their characters. In Egypt, for example, the hieroglyphic, hieratic and demotic writing systems, all of which consisted of a mixture of logograms (or ideograms), consonantal phonograms and semographic determinatives (see, for example, Ritner 1996a), employed specific characters for writing numbers. However, as soon as the Coptic alphabet is developed in the early Roman period from the Greek alphabet with the addition of six or seven uniconsonantal demotic signs (cf., for instance, Ritner 1996b), employing letters for numerical notation becomes very easy and this practice is then used widely in Coptic on the model of the Greek system. The same is the case with Syriac, even though there the influence of the Greek alphabet is less obvious and direct than on Coptic (Healey 1990, 60; Al-Jadir 2006 [2009], 7–8). Clearly, the influence of Greek literate culture had become so pervasive in the eastern Mediterranean by the Roman period that most, if not all, scribes of other languages were familiar with Greek practices as well and adopted these wherever the possibility and need arose.

Cryptography

The next secondary function of alphabetic scripts of the ancient eastern Mediterranean that I would like to mention and briefly describe here is that of cryptography or cipher, that is, a secret system of writing that prevents the uninitiated from understanding the text. Again, Greek supplies an excellent example, an example that had a great deal of influence on a variety of other alphabetic scripts in the ancient Near East. This cipher is often referred to as the ΑΘΒΗ system. It is based on the numerical value of the letters of the Greek alphabet assigned to these according to the alphabetic system of numerical notation that I have just described. As we have seen, under this system the 24 letters of the Greek alphabet plus three archaic signs were each given a numerical value ranging from 1 to 900 that corresponded to their order in the Greek alphabet.

The ΑΘΒΗ cipher was created by dividing the 24 letters of the Greek alphabet plus the three archaic signs that the alphabetic system of numerical notation included into three groups of nine letters each and then by assigning to these the letters of each group in the reverse order.⁶ Thus, within the first group of alpha to theta, alpha was paired with theta, beta with eta – hence the name of this cipher – and so on. The interesting thing is that the total of the numerical value of each pair of letters in the first group of nine letters is 10. The total of the numerical value of each pair of letters in the second group of nine letters is 100, while that in the third and final such group is 1000. Thus, for a perfect such cipher, one needs 27 letters. It is interesting to note that the central letter of each group: epsilon of the first, nu of the second and phi of the third, is each paired with itself, in other words, they are left unsubstituted in the cipher as the double of the numerical value of each yields the required total, 10 in the case of the two epsilons, 100 in the case of the two nus and 1000 in the case of the two phis. Thus, in this cipher that this code creates, these three letters, that is, epsilon, nu and phi, remain undisguised. Although this system requires and is based on 27 letters, three groups of nine, in actual fact only the 24 letters of the Greek alphabet are used in the code as the archaic letters of stigma (or digamma), qoppa and sampi were not normally used in late classical and later Greek texts.

The encoding worked by the scribe writing not the letters required but their corresponding pairs according to the cipher. This method of encoding created a simple secret writing system that enjoyed a great deal of popularity in antiquity and the Middle Ages. Its earliest attestation to date is found in a visitor’s graffito carved into the Memnon colossus in Western Thebes and dated most probably to the second century AD although an earlier date for it cannot be ruled out completely. The cipher remained popular among Byzantine scribes and is attested as late as the mid-sixteenth century. In antiquity it appears to have been used in graffiti, a magical papyrus but even in an ordinary private list of clothes and victuals. Some scholars have recently speculated that the cipher might already have been known to the traveller and versatile intellectual Sextus Julius Africanus, who lived approximately between AD 160 and 240. In the Byzantine period it was used by scribes of manuscripts for writing colophons.

The popularity of this cipher in Greek is also reflected by the fact that it was adopted by scribes of other languages as well quite early on. The same idea of coding underlies a popular Syriac cipher that is attributed to the Parthian sage Bardaisan, a contemporary of Sextus Julius Africanus, and is often referred to as the ‘Alphabet of Bardaisan’ (Paz and Weiss 2015, 51–54, 63–65). The two intellectuals personally met in the court of King Abgar the Great of Edessa and so it is possible that the basic idea of this cipher was transmitted to Bardaisan there, and then on to Syriac scribes. Be that as it may, Syriac scribes used the cipher widely for the same purpose as Byzantine scribes, that is, writing colophons at the end of manuscripts they copied. The practice in Syriac is documented to have continued even later than in Greek, with one of the latest Syriac examples being found at the end of a manuscript dated to 1831.

The idea of the Greek ΑΘΒΗ cipher was also adopted into Coptic (Paz and Weiss 2015, 59–65). So far the earliest known attestation of this Greek cipher is found in Nag Hammadi Codex VIII, below the subscript title of the first tractate, entitled Zostrianos, contained by the codex. The codex is dated to the fourth century AD. What is surprising about this three-line cryptogram is that, although it